The proposed work will provide a substantial amount of much needed data on the behavioral pharmacology of cocaine. This drug will be compared to another widely abused "stimulant", amphetamine, in both learning and performance situations with two species (pigeons and monkeys). Both acute and chronic effects will be assessed. More specifically, in the experiments on learning, the subjects will obtain food by making four responses on three keys in a certain sequence, e.g., Left, Center, Right, Left. Learning is defined by the decrease in errors across trials within a session. To establish a baseline of repeated acquisition, the sequence of correct responses is changed from session to session. For comparison, the drugs will also be studied under a "performance" condition, in which the sequence of correct responses is the same from session to session.